Abstract: To address the complexities and subjectivity in traditional rock mass classification methods for drill-andblast tunnels, which have limitations when assessing complex rock masses, this study proposes an intelligent method for identifying tunnel surrounding rock grades. The method involves fitting the distribution of drilling parameters using the kernel density estimation method and employing a Naive Bayes classification algorithm for rock classification.The performance of the classification model was enhanced through cross-validation, and the method was validated using literature data. The results demonstrate that the Naive Bayes classification algorithm based on kernel density estimation can accurately classify the quality of the tunnel face surrounding rock using drilling parameters, achieving a classification accuracy of 94.0% in the test set. Furthermore, the cross-validation method improved the model′s performance, reaching a classification accuracy of 98.7% on a test set of 299 samples.